Abstract: For large-scale spatial data sets, the computational efficiency of the current serial topology checking algorithms have reached their limit. More computing resources are needed to achieve efficient parallel computing of topological relation. Based on the analysis of hierarchical and topological features of topological relational computing, a parallel computing method based on vector data partition and Q&R parallel index was proposed in this paper. A parallel topological relationship computing middleware was developed and deployed in the cloud environment, and applied to the quality checking of cadastral data mass. The application results in the virtualized cluster resources of cloud environment demonstrate that, the proposed method can realize the high balance of computing task load and the basic balance of data load among the process; the acceleration ratio is linearly positively correlated with the number of processes, and the computing efficiency of parallel topological algorithm is stable at 80%. This paper presented an efficient and available parallel computing algorithm and parallel computing middleware in the cloud environment for topological relationship quality checking service of massive cadastral databases in various high-performance computing environments.

Key words: topological relationship, parallel computing, cloud computing, data partition, load balancing